JOURNAL OF GEOPHYSICAL RESEARCH, VOL 88, PAGES 493-499, JANUARY 1, 1983
A THRESHOLD EFFECT FOR SPACECRAFT CHARGING
Richard Christopher Olsen
Physics Department, The University of Alabama in Huntsville, Huntsville, Alabama 35899
Abstract. The existence of a threshold energy dependence is
shown for the charging of geostationary spacecraft in eclipse.
Applied Technology Satellite 6 and P78-2 (SCATHA) data show
that plasma sheet fluxes must extend above 10 keV for these
satellites to charge in eclipse. The existence of a threshold
energy is significant because magnetospheric convection
boundaries produce relatively sharp phase space boundaries,
with electron fluxes falling rapidly above the energy associated
with the local Alfven layer. The threshold effect is due to the
shape of the normal secondary yield curve, in particular the high
energy crossover, where the secondary yield drops below 1. A
large portion of the ambient electron flux must exceed this
energy for a negative current to exist.
Introduction
Eclipse charging events on geosynchronous satellites have
been explained as the natural result of the balance between the
currents due to ambient electrons, ambient ions, and the
secondary fluxes generated by the ambient plasma [DeForest,
1972; Garrett and DeForest, 1979]. This paper presents the
borderline case between no charging and large (kV) negative
potentials, and the dependence of this transition on a threshold
energy in the ambient plasma. The motivation for this study is
the desire to understand the observation that in hundreds of
eclipses observed by Applied Technology Satellites 5 and 6 and
P78-2 (SCATHA), the observation of large negative potentials
coincided with substantial fluxes over 10 keV. Also, in a few
percent of the eclipse events observed on these satellites,
relatively subtle changes in the environment produced substantial
changes in the spacecraft potential. These changes in the
environment were generally due to the convection of stable
plasma populations past the satellite or motion of the spacecraft
with respect to convection boundaries, rather than injections of
hot plasma (substorms). (See Kivelson et al. [1979] for
examples.) Of course, injections due to substorms can also
provide the necessary increase in net current. In the following
sections, the basic theory for eclipse charging is reviewed,
followed by a statistical view of the charging dependence on an
electron threshold energy, based on a survey of spectrograms.
Next, an example of the changes in potential caused by small
environment changes is shown, and then a recap of the statistical
survey is shown, using a different criterion for the threshold
energy.
Copyright 1983 by the American Geophysical Union.
Charging Theory
The net current to a spacecraft in eclipse can be symbolized
I net = q ( Fi - Fe + Fsec )
Fi.
= ion flux integrated over the spacecraft
surface;
Fe = electron flux integrated over the
spacecraft surface;
Fsec = secondary fluxes due to primary impact.
In equilibrium, the net current to the spacecraft is zero.
Though secondary electron production due to ion impact is
important, the major secondary yield term is that due to electron
impact. Figure 1 shows a typical secondary yield curve, in this
case for kapton, a common spacecraft material (adapted from
Leung et al. [1981]). This curve shows the features typical of
most materials. There is a peak yield, greater than 1, at energies
of a few hundred electron volts. If most of the ambient flux is in
this energy range, the outgoing secondary flux will exceed the
incoming electron flux, and the spacecraft will charge positively
until the outgoing fluxes are limited sufficiently to attain
equilibrium (typically a few volts positive; see Olsen [1982]).
Other common features are lower and upper crossover energies,
where the yield equals 1. It is this characteristic of the yield
curve which leads to the threshold energy effect. As the energy
of the ambient plasma population increases, the net current due
to electrons will eventually become negative, and the potential
will then be determined by the balance of these fluxes and the
ambient ion fluxes, at some negative potential. This
characteristic of the secondary yield curve has been recognized,
and indeed utilized, for laboratory measurements (see Leung et
al. [1981], for example). Rubin et al. [1978] have used this
characteristic of the secondary yield curve to show that the
temperature of a Maxwellian distribution would determine
whether or not charging would occur, and that there is a critical
temperature for charging. The upper crossover energy of 1.0 keV
shown here for kapton, and the maximum yield of 1.8, are
relatively low compared to some of the other dielectrics
encountered on spacecraft, such as teflon and mylar [Willis and
Skinner, 1973], so it should be kept in mind that a spacecraft
may require electron fluxes at well over 1 keV in order to charge.
The dependence of surface potential on material properties has
been explored by Knott [1972] and Prokopenko and Laframboise
[1980], among others, using a few typical particle spectra. In the
example which follows, the complementary case is presented of
a constant material (the spacecraft) being exposed to a varying
environment.
Paper number 2A1406.
0148-0227/83/002A-1406$02.00
493
JOURNAL OF GEOPHYSICAL RESEARCH, VOL 88, PAGES 493-499, JANUARY 1, 1983
Fig. 1.
Secondary electron yield due to electron impact for
kapton. The number of secondary electrons produced by electron
impact is plotted as a function of the energy of the primary
electron [Leung et al., 1981].
quantified. The survey was conducted by looking at
spectrograms, as shown in Figure 2. The normal features of an
eclipse charging event are found in these data from ATS 6 on
October 4, 1974. This grey scale presentation of the particle
fluxes uses black for zero flux, white for high fluxes, and grey
for intermediate values. Plasma sheet electrons are visible
throughout the 4-hour period shown here, extending from a few
hundred eV to 10-20 keV. High photoelectron fluxes from 1 to
15 eV disappear during the eclipse (0530-0640 UT), with
negative potentials evident from 0545 to 0640 UT in the ion
data. Ambient ions reaching the spacecraft are accelerated by the
spacecraft potential, and the absence of locally generated ions
provides the characteristic charging signature of a large black
region in the ion plot. There is no obvious differential charging
signature in this data set [Olsen, et al. 1981; Whipple, 1976].
The transition energy is defined as the point where the
electron fluxes drop from white to black (about 15 keV). This is
partly an artifact of the logarithmic compression applied to the
data before plotting, and inferences of boundaries in phase space
cannot be safely done without creating and studying distribution
functions. Since the count rate is proportional to E2f, where f is
the distribution function, the count rate naturally begins to drop
above the normal peak at twice the temperature (for a
Two relatively new terms are introduced below, i.e., transition
energy and threshold energy. The first is the energy where the
particle flux drops rapidly, or the energy where there is a distinct
change in the distribution function. The term threshold energy is
used to describe the transition energy associated with the
beginning of the charging process.
Spacecraft and Instruments
Applied Technology Satellite 6 (ATS 6) was launched in
1974 into a geosynchronous orbit, where the three-axis-stabilized
satellite con-ducted a variety of engineering and scientific
experiments. The UCSD auroral particles experiment was an
electrostatic analyzer (ESA) designed to measure ion and
electron fluxes from 1 eV to 80 keV [Mauk and Mcllwain, 1975].
Two sets of electron and ion detectors sampled particle fluxes by
rotating in orthogonal planes, providing good pitch angle and
energy coverage in a few minutes. Energy resolutions (∆E/E) is
20%, angular resolution 5° x 7°. The east-west (EW) detector
assembly suffered a partial failure prior to the first eclipse
season, limiting the energy range covered by this head. It still
provides useful data within the energy range covered.
P78-2 (SCATHA) is an Air Force satellite designed to study
spacecraft charging at and near geosynchronous orbit. It is in an
equatorial, elliptical orbit, with perigee at 5.5 RE. and apogee at
7.5 RE. It was spin stabilized, with a spin period of about 57 s. A
slightly modified version of the UCSD experiment from ATS 6
was flown on SCATHA.
Charging Survey
The initial motivation for this work was the observation that
charging on ATS 6 and SCATHA required substantial fluxes
over 10 keV. In this first view of the data, this observation is
Fig. 2. ATS 6 energy-time spectrogram for October 4, 1974. Four
hours of data from the NS detector show the ion and electron data
around the eclipse event at local midnight. High fluxes are white,
low fluxes are black. The energy scale is roughly logarithmic,
going from 1 eV to 80 keV. Periodic vertical striping is due to
detector rotation and pitch angle variations in the ambient plasma.
494
JOURNAL OF GEOPHYSICAL RESEARCH, VOL 88, PAGES 493-499, JANUARY 1, 1983
Maxwellian). This initial survey should be viewed with some
caution, therefore, since the transition energy used in Figure 3
can be an artifact and not necessarily the convection boundary
demonstrated in subsequent portions of this paper.
Figure 3 is a plot of observed eclipse potentials against
transition energies, as determined from spectrograms. Again, the
transition energy is the point where the electron count rate drops
abruptly, typically from 10,000 Hz to a few hundred hertz. ATS
6 data from 40 eclipses in the fall of 1974 are combined with
SCATHA data from 80 eclipses in 1979. Both spacecraft show a
sharp increase in charging events when the electron flux goes
above 15 keV. Once the threshold is crossed, however, the
potentials reached by the spacecraft depend on other details of
the ion and electron distribution functions, and no clear trend is
found relating potential and threshold energy. In some of the
largest ATS 6 and SCATHA charging events, the transition
energy is over 80 keV.
FIGURE 3. ATS 6 and SCATHA charging dependence on an
electron threshold energy. The threshold energy used here is
essentially the point where the electron count rate drops below the
1000 Hz point. This generally corresponds to the upper energy
cutoff of the plasma sheet at the spacecraft location but can
simply be an artifact of the detector response to keV particle
distributions.
October 4, 1974
Figure 2 shows the event selected for more detailed analysis
of the threshold energy and the relationship of this energy to
variations in the electron distribution function and phase space
boundaries. This is one of two events in the fall 1974 ATS 6 data
set, where relatively small changes in the environment produced
substantial changes in the spacecraft potential. Figure 4 shows
the count rate for ions and electrons perpendicular to the
magnetic field prior to and after the spacecraft charges.
(Variations in the distribution function with pitch angle can be
important, but in this case, the lack of recent magnetic activity
has allowed the field-aligned portion of the electron population
to be lost. The majority of the current is in the perpendicular
fluxes.) The NS detector is rotating, sampling the pitch angle
distribution of the plasma, while the EW detector is parked,
measuring fluxes perpendicular to the magnetic field. The
spacecraft was initially uncharged in eclipse (i.e., a few volts
positive) and then, following a slight increase in electron energy
at 0545 UT, began to charge negatively, reaching a maximum
potential of.-700 V. This progression of events is visible in
Figure 2 and will be expanded on in Figure 6.
Figures 4a and 4b show the ion and electron count rates from
the NS and EW detectors at 0536 UT, when the spacecraft is
near zero volts potential. The ion data show the absence of
measurable potential, as was noted in the spectrogram. The
minimum in ion flux at 20 keV is the 'deep proton minimum' as
described by Mcllwain [1972]. This feature is the boundary
between eastward and westward convecting ions, and the
constancy of this feature in this example is evidence for a lack of
substorm activity near the spacecraft. The electron data show a
peak at about 7 keV and a rapid drop beginning at about 9 keV.
The differences between the NS and EW detectors apparent in
Figure 4b are due to pitch angle variations in the ambient plasma
and differences in spiraltron efficiencies. Field-aligned data (not
shown) reveal a normal electron loss cone. The 10 eV to 1 keV
ion data show a source cone distribution.
Figure 4c and 4d show the ion and electron data at 0557 UT,
when the spacecraft has reached a potential of -650 V. The peak
in the ion flux between 600 and 700 eV shows the spacecraft
potential signature. The energy of the proton minimum has
dropped slightly during this period as a result of the spacecraft
motion. These changes in the ion flux at high energy do not
greatly affect the spacecraft current balance, and the ion data are
ignored for the rest of the paper. The electron data show that the
flux peak has moved up to 8 or 9 keV, with the drop in flux
beginning at 12 keV.
Figure 5 shows the electron distribution functions for the NS
detector data just shown, along with the result of fitting the 2- to
8-keV electron data with a Maxwellian. The 5-keV fits can be
overlapped by shifting the distribution function by 650 eV, the
change in spacecraft potential. This implies that other differences
in the distribution functions are due to environmental changes.
495
JOURNAL OF GEOPHYSICAL RESEARCH, VOL 88, PAGES 493-499, JANUARY 1, 1983
FIGURE 4. Electron and ion count rates from the NS and EW detectors on October 4, 1974. (a) Ion count rates at 0536 UT. One energy
scan is plotted from the NS detector, while three scans have been averaged for the EW plot. Since the accumulation time is 4 s, count
rates below 100 Hz are disturbed by statistical fluctuations. (b) Electron count rates at 0536 UT, one scan from each detector. NS fluxes
are plotted as a solid line, EW fluxes as crosses. (c) Ion count rates at 0557 UT. One scan is plotted for the NS detector, two are averaged
for the EW plot. (d) Electron count rates at 0557 UT, as in Figure 4c.
Also, the drop in the count rate at 9 keV at 0536 UT can be seen
as the point where the observed distribution function drops
below the nominal value it should have for a 5-keV Maxwellian.
Similarly, the 0557 UT data begin to drop below the distribution
function fit at 12-14 keV. This bend or break in the distribution
function implies there is a phase space boundary at that energy.
A phase space boundary is a boundary in velocity space for
particle distribution functions. They are useful for separating
plasmas with different origins, particularly if the populations are
approximately Maxwellian, as electron distributions often are
[Whipple, 1976; Greenspan et al. 1981]. In this case, the fact that
the same temperature fit can be made for the 3- to 7-keV
electrons implies that the nature of the plasma has not changed
(i.e., it has not been heated). The changes in the distribution
functions at 9 keV (0536 UT) and 12 keV (0557 UT) are
apparently due to the Alfven boundary, the result of
magnetospheric convection electric fields. The changes in the
electron flux are therefore attributed to the crossing of the 10keV Alfven boundary.
The changes in the spacecraft potential are therefore due to
changes in the electron flux above 9 or 10 keV and not to the
temperature of the distribution. Figure 6 shows this relation-ship
over the time period when the potential was changing from 0 to 700 V, along with the electron fluxes at 6.5, 8.7, 10,0, and 11.5
keV. The lowest energy count rate shows little change over the
time period and reflects the constancy of the plasma sheet
population. The higher energy electron count rates show the
changes attributed to the crossing of magnetospheric boundaries.
The most dramatic change is in the 10-keV electron flux, which
is initially excluded from the region of the spacecraft. Electrons
of this energy associated with the 5-keV plasma population are
on drift trajectories which are outside the initial spacecraft
position. Changes in the spacecraft position (and perhaps the
drift boundary) over the period of the eclipse bring the spacecraft
into the 10- to 15-keV fluxes. The 8.7- and 10.0-keV fluxes both
begin their climb between 0540 and 0545 UT, which is also the
time associated with the beginning of the charging period. The
spacecraft potential is deter-mined from the NS ion detector, and
substantial fluxes are not measured in a charging peak until the
potential has reached the -20 V point. This is largely a result of
the lack of ambient low energy ions, and extrapolating the curve
back to the -10 V point shows that the charging probably began
at about 0540 UT, the time for the beginning of the increase in 9to 10-keV electron fluxes.
496
JOURNAL OF GEOPHYSICAL RESEARCH, VOL 88, PAGES 493-499, JANUARY 1, 1983
The change in the current to the spacecraft can be calculated
from the observed changes in the distribution function. The
observed change in the primary electron current density is about
15 pA/cm2. Since the previously trapped electrons are now
escaping and the spacecraft is nominally still near equilibrium
(zero net current), this value is comparable to the in-crease in the
current due to escaping secondary electrons. The increase in ion
current is largely contained in the ion charging peak. The current
density in the peak is only 1 pA/cm2.
Statistical Look 2
Fig. 5. Electron distribution functions for October 4, 1974, at 0536
and 0557 UT. The earlier time is plotted as a solid line, the latter
as crosses. The parallel diagonal lines are the results of least
squares fits between 3 and 7 keV. These fits gave electron
-3
temperatures of 5 keV and a density of about 1 cm .
Figure 7 shows a subset of the data from Figure 3, with the
data points selected from those with a transition energy near 10
keV. In this plot, threshold energies are determined from
distribution function plots for those days where it is clear that
there is a distinct drop in the electron distribution function and
therefore a boundary in phase space. The transition energy has
been corrected for the spacecraft potential, so these- boundaries
are those for the ambient plasma. It is clear that the crucial
energy is about 15 keV for the cases shown here. The line noted
D266 is a sequence of data for an ATS 6 event similar to the one
just shown. Again, once the threshold has been passed, the
correlation between threshold energy and potential is not a clear
one.
Fig. 6. Electron flux and spacecraft potential on October 4, 1974, in eclipse. Fluxes from the EW detector at 6.5, 8.7, 10.0, and 11.5 keV
are plotted versus time here, along with the potential determined from the ion charging peak. The potentials between 0545 and 0555 UT
were least square fitted for the 3.85 min time constant shown plotted through the potentials.
497
JOURNAL OF GEOPHYSICAL RESEARCH, VOL 88, PAGES 493-499, JANUARY 1, 1983
correlated the spacecraft potential with particle fluxes in the 30to 80-keV energy range. This differs slightly from the nature of
the findings here, in that the threshold effect determines if the
spacecraft will charge, not the magnitude of the charging. The
same comment applies to distinguishing between the correlation
established between 25-. to 75-keV energy electrons and
charging by Gussenhoven and Mullen [1982]. One other
distinction between this work and that of Gussenhoven and
Mullen is that the correlation they show is for daylight charging,
which involves different physical processes. Note also that their
next lower channel, nominally from 11 to 35 keV, will probably
sample electrons below the threshold energy, thus obscuring the
importance of this energy range.
The threshold energy of 10 keV found here will vary from
spacecraft to spacecraft, but the similarity of the responses by
ATS 6 and SCATHA, which are vastly different spacecraft,
suggests that any attempt to monitor the magnetosphere for
charging prediction should include the 10- to 20-keV energy
range.
Acknowledgments. The data presented here were provided
by C. E. Mcllwain and analyzed under NASA contract NAS833982.
The editor thanks H. B. Garrett and J. B. Reagan for their
assistance in evaluating this paper.
References
FIGURE 7. Spacecraft potentials are plotted against threshold
energies determined from distribution functions.
Conclusions
The data from ATS 6 and SCATHA shown here have
demonstrated the dependence of charging on the electron fluxes
between 10 and 20 keV.
The transition from small positive potentials to large
negative potentials has been shown to be associated with the
movement of the spacecraft into plasma sheet regions `outside
the 10-keV Alfven layer. Even in cases where the distribution
function does not show characteristics associated with these
spatial boundaries, it appears that the spacecraft does not charge
unless there are substantial fluxes between 10 and 20 keV. For
the UCSD detectors, this means the differential energy flux in
this energy range should be above 108/s cm2 or a current density
of 10 pA/cm2. This criterion for charging is useful for those
concerned with predicting charging effects, particularly from the
viewpoint of spacecraft safety.
The concept of a critical energy range has been noted by
Reagan et al. [1981], who found that for SCATHA, the critical
energy range was between 3 and 30 keV. This result remains
valid in the context of the work shown here; the critical energy
range has simply been narrowed.
The idea of a threshold dependence has been noted
experimentally for ATS 5 and ATS 6, again by using the UCSD
data sets, by Garrett [1981] (see his Figure 19). Garrett found
that charging required an electron temperature of at least 2 keV
(determined by integrating the distribution function and taking
the ratios of the moments). The main difference in this study is in
the identification of the critical energy range which causes the
temperature to rise to a point where charging can begin. Garrett
et al. [1980] have studied charging events for ATS 6 and
DeForest, S. E., Spacecraft charging at synchronous orbit, J.
Geophys. Res., 77, 651-659, 1972.
Garrett, H. B., The charging of spacecraft surfaces, Rev.
Geophys. Space Phys., 19, 577-616, 1981.
Garrett, H. B., and S. E. DeForest, Time-varying photoelectron
flux effects on spacecraft potential at geosynchronous orbit,
J. Geophys. Res., 84, 2083-2088, 1979.
Garrett, H. B., D. C. Schwank, P. R. Higbie, and D. N. Baker,
Comparison between the 30- to 80-keV electron channels
on ATS 6 and 1976-059A during conjunction and
application to spacecraft charging prediction, J. Geophys.
Res., 85, 1155-1162, 1980.
Greenspan, M. E., M. B. Silevitch, and E. C. Whipple, On the
use of electron data to infer the structure of parallel electric
fields, J. Geophys. Res., 86, 2175-2182, 1981.
Gussenhoven, M. S., and E. G. Mullen, A 'worst case' spacecraft
charging environment as observed on SCATHA on 24 April
1979, paper presented at 20th Aerospace Sciences Meeting,
Am. Inst. of Aeronaut. and Astronaut., Orlando, FL., 1982.
Kivelson, M. G., S. M. Kaye, and D. J. Southwood, The physics
of plasma injection events, in Dynamics of the
Magnetosphere, edited by S.-I. Akasofu, pp. 385-405, D.
Reidel, Hingham, MA., 1979.
Knott, K., The equilibrium potential of a magnetospheric satellite
in an eclipse situation, Planet. Space Sci., 20, 1137-1146,
1972.
Leung, M. S., M. B. Tueling, and E. R. Schnauss, Effects of
secondary electron emission on charging, Spacecraft
Charging Technology 1980, edited by N. J. Stevens and C.
P. Pike, NASA Conf. Publ., 2182, 163-178, 1981.
Mauk, B. H., and C. E. Mcllwain, ATS 6 UCSD auroral particles
experiment, IEEE
Trans. Aerosp. Electron. Syst.,
AES-11, 1125-1130, 1975.
Mcllwain, C. E., Plasma convection in the vicinity of the
geosynchronous orbit, in Earth's Magnetospheric Processes,
498
JOURNAL OF GEOPHYSICAL RESEARCH, VOL 88, PAGES 493-499, JANUARY 1, 1983
edited by B. M. McCormac, pp. 268-279, D. Reidel,
Hingham, MA., 1972.
Olsen, R. C., The hidden ion population of the magnetosphere, J.
Geophys. Res., 87, 3481-3488, 1982.
Olsen, R. C., C. E. Mcllwain, and E. C. Whipple, Observations
of differential charging effects on ATS 6, J. Geophys. Res.,
86, 6809-6819, 1981. Prokopenko, S. M. L., and J. G.
Laframboise, Differential charging of geostationary
spacecraft, J. Geophys. Res., 85, 4125-4131, 1980.
Reagan, J. B., R. W. Nightingale, E. E. Gaines, R. E. Meyerott,
and W. L. Imhof, Role of energetic particles in
charging/discharging of spacecraft dielectrics, Spacecraft
Charging Technology 1980, edited by N. J. Stevens and C.
P. Pike, NASA Conf. Publ., 2182, 74-85, 1981.
Rubin, A. G., P. L. Rothwell, and G. K. Yates, Reduction of
spacecraft charging using highly emissive surface materials,
in Proceedings of the
1978
Symposium
on
the
Effects
of
the Ionosphere on Space and
Terrestrial Systems, edited by John M. Goodman, pp. 313316, Naval Research Laboratory, Washington, D. C., 1978.
Whipple, E. C., Jr., Observation of photo-electrons and
secondary electrons reflected from a potential barrier in the
vicinity of ATS 6, J. Geophys. Res., 81, 715-719, 1976.
Willis, R. F., and D. K. Skinner, Secondary electron emission
yield behavior of polymers, Solid State Commun., 13, 685688, 1973.
(Received May 5, 1982;
revised July 26, 1982;
accepted September 13, 1982.)
499